Scholarlink

The paper presents a new approach to Asset Management that utilizes Industry 4.0 techniques and Artificial Intelligence (AI), specifically Machine Learning (ML). The research emphasizes early anomaly detection and failure mode classification in operating equipment in dynamic ranges. The study is based on the secondary cryogenic pumps of a Liquefied Natural Gas (LNG) regasification plant, which operate in a highly dynamic environment with rapid changes in pressure. Traditional condition monitoring methods, which rely on monitoring operational parameters and physical measurements, are contrasted with the proposed ML-based approach. ML models aim to detect early-stage issues and forecast critical failures by exploiting complex and nonlinear relationships among multiple operational variables. For the study, six years of historical data gathered by the different sensors of the 4 pumps have been used.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Early Failure Detection in Secondary Cryogenic Pumps Through Machine Learning Techniques in the Context of Industry 4.0

  • 章节

  • 发表日期：

    2026年

查看全文
  • Sonia Liñán García,
  • Antonio de la Fuente Carmona,
  • Javier Serra Parajes,
  • Adolfo Crespo Márquez

摘要

The paper presents a new approach to Asset Management that utilizes Industry 4.0 techniques and Artificial Intelligence (AI), specifically Machine Learning (ML). The research emphasizes early anomaly detection and failure mode classification in operating equipment in dynamic ranges. The study is based on the secondary cryogenic pumps of a Liquefied Natural Gas (LNG) regasification plant, which operate in a highly dynamic environment with rapid changes in pressure. Traditional condition monitoring methods, which rely on monitoring operational parameters and physical measurements, are contrasted with the proposed ML-based approach. ML models aim to detect early-stage issues and forecast critical failures by exploiting complex and nonlinear relationships among multiple operational variables. For the study, six years of historical data gathered by the different sensors of the 4 pumps have been used.